Van Til believed that only the Christian worldview provides the metaphysical precondition for the possibility of knowledge on the grounds that it provides an answer on how to relate the diversity of experience to a unified ground of knowledge. This precondition for intelligibility lies in God's Triune nature. However, it is impossible for man to understand God's nature. Therefore, it is impossible for man to understand the precondition for knowledge. Since man cannot understand how finite knowledge relates to the metaphysical precondition for knowledge, and since only God possesses this knowledge, man can never know in exactly the same manner as God.
However, this is not the only respect in which Van Til believed that God's knowledge is qualitatively different from ours. He also held that God alone knows how all of the propositional facts in the universe relate to one another, and held that in this sense, God's knowledge was qualitatively different from our finite knowledge. It is on these grounds that he held that knowledge must be analogical rather than univocal. Gordon Clark ably explains the genesis of this distinction in philosophy:
The notion of analogy begins quite simply and innocently in Aristotle. He notes that when we call a book a medical book, and when we call an instrument a medical instrument, and when we call a man a medical man, the predicate medical does not bear exactly the same sense in the three instances. The term is not equivocal, as is the case when we call Argos the dog of Ulysses and when we call Sirius the dog in the sky; but on the other hand, the term is not strictly univocal. It is analogical.
This simple distinction was elaborated by the Scholastic and the Neoscholastics into a complicated theory, in which, it would seem, the original situation no longer serves as a solid basis. The motivation and intricacies of the theory are seen most clearly in the arguments for the existence of God and our knowledge of him. God, according to the Thomists, is an absolutely simple being: but a simple, eternal, and immaterial being cannot constitute an object proportionate to our human understanding. Simplicity and eternity are not factors in our world of experience, and therefore we have no positive concept of them. To say that God is eternal means nothing more than that God is not temporal. What eternity positively means remains unknown to the human mind. What man has in this instance may be called negative knowledge.
Similarly, when we call God wise and when we call a man wise, the term does not bear the same sense. God’s wisdom is not distinct from his essence or his being; but the wisdom of man is. In general, there is no affirmation whatever that can be made of God and of man in the same sense. The reason for this impossibility is not only that the predicates do not bear the same meaning in both cases, but that, far more radically, the copula is bears two different senses. In God essence and existence are identical: What God is and that God is are the same. In every case other than God this is not so. Accordingly, when we say God exists and when we say man or dog exists, the term exist does not mean the same thing. Therefore, no term, not even the copula, can be used univocally of God and man.
Now, if the only alternative to univocal predication were equivocal predication, knowledge of God derived by abstraction from experience would be patently impossible. When words are used equivocally there is no definite relationship between the meanings, and knowledge of God would be in a state similar to a knowledge of Sirius that would be based on an experience of Ulysses’ dog. To avoid this fatal difficulty, the Thomists are forced to find some intermediate between univocal and equivocal predication, and they appeal to analogy. Between Argos and Sirius there is no resemblance, but in the case of God, man resembles God, they say, though God does not resemble man. This resemblance permits us to attach some meaning to the statement God is, so that we are neither in complete ignorance, nor limited to negative knowledge, but have an analogical if not a univocal knowledge. Thus empiricism in its Thomistic form attempts to escape the limits of experience.
There seems to be a very serious objection to this theory of analogy. Aristotle’s original analogies cause no difficulty. The term medical, whether applied to a man, a book, or an instrument, is presumably derived from experience. In all three cases there is a relationship to the science of medicine. And for this reason there is a univocal basis for the analogy. The term medical might univocally be defined as “having to do with the science of medicine”; and in this univocal sense the man, the book, and the instrument are all medical. Similarly, all the analogies of common speech have a univocal basis. The paddle of a canoe is analogous to the paddles of a paddle-wheel steam boat; it may even be said to be analogous to a screw propeller. It is so because there is an area of common or univocal meaning. The paddle and the screw propeller are both devices for using power to make boats move through the water. The Neoscholastics list and classify different types of analogy; some are more complicated than the preceding. For example, it might be said that the mind is to the soul as the eye is to the body. Here there is analogy, possibly between the mind and the eye, or possibly between two relationships. But no matter how complicated, or what type of analogy, an examination must discover some univocal element. The two terms must be like each other in some respect. If there were no likeness or similarity of any sort, there could be no analogy. And the point of likeness can be designated by a simple univocal term or phrase. The Thomists admit the likeness or resemblance in analogy; they deny the univocal basis. They transfer analogy from the status of a literary embellishment or pedagogical aid to that of a serious epistemological method. But this removes every real distinction between analogy and equivocation. (Gordon Clark, A Christian View of Men and Things)
So much for the historic, neo-Aristotelian Thomist motivation for the development of analogical reasoning. It is clear that Aquinas' religious metaphysics constitute his motivation for his theory of analogical knowledge and predication. While Van Til attempts to distance himself from the scholastic Aquinas, Van Til's own motivations for advocating his distinct brand of analogical knowledge are strikingly similar to that of Aquinas. Like Aquinas, Van Til believed that the incomprehenesibility of God necessitated an analogical theory of knowledge on the ground that God's Triunity, though it constituted the precondition of intelligibility and knowledge, was incomprehensible to man. Indeed, for Van Til, we have no familiarity with anything like Triunity this side of the creaturely realm; it is purely an incommunicable and incomprehensible attribute of the radically transcendent Creator-God. It is this radical incommensurability with anything with which we are ordinarily familiar which, for Van Til, necessitated an analogical theory of knowledge. But there is another reason Van Til believed that knowledge and predication of God had to be analogical:
...Van Til insisted upon the fact that to have truth in one's mind that mind must be built upon other propositions. The truthfulness or falsity demanded that the individual proposition be held in the midst of certain other basic propositions that must be consciously present in that mind in order to correctly know truth. Now, of course, God knows every proposition in the context of all other propositions for Van Til, and, therefore, the limited human mind never knows it the way God does. Van Til had an expression, of repeated: "true as far as it goes," meaning, of course, that for that mind which holds all propositions in a system, the more complete the system, the more full the truth. With growth in the knowledge of basic propositions, the further than mind had the truth. Van Til's concept is that for relative human beings, they can have all needful truth but never perceive it as God does with his infinite knowledge of everything that affects any proposition. He charged Clark, therefore, with denying the incomprehensibility of God and Clark charged him with agnosticism since...for him it was impossible to know anything as God did. Clark wanted an absolute even if it were only in the single proposition. (Gordon Clark: Personal Recollections, pgs. 103-104)
Thus, Van Til's motivations for holding to analogical knowledge are twofold:
1) God's Triunity resolves the perennial philosophical problem of how to relate the diversity of experience to a unified ground, though this solution is itself incomprehensible to man. Since it is incomprehensible, we cannot understand how finite knoweldge can relate to this unified ground. Only God understands His Triunity, and therefore only God understands how all knowledge relates back to this Triune preecondition of intelligibility. Therefore, God understands all things in a qualitatively different manner than man.
2) Van Til and Clark believed that all propositions only exist within the sum total of all propositions. Since only God understands the relation each proposition to every other proposition, Van Til believed that, given any proposition, we could never know a proposition the way God does. Analogical knowledge is therefore required. Clark, of course, believed that each proposition exists in a context of all other propositions, and that only God understands the relation it bears to all others, but he did not believed this necessitated belief in analogical knowledge. It is perfectly possible to abstract individual propositions from their context and acquire partial knowledge by understanding them in isolation.
In a conversation with my good friend Joel Tay, he pointed out that this position of Van Til's is self-refuting. Suppose we apply Van Til's own standard of analogical knowledge to the very notion of analogical knowledge. Ironically, an univocal definition of analogical knowledge is required in order to even know what analogical knowledge is. In other words, if one is told that we must hold to an analogical theory of knowledge, one must have a working definition of analogical knowledge.
This definition must not itself be analogical. If it is itself merely an analogy, we simply do not know what "analogical knowledge" is. If such a thing cannot be provided, then we are not left with analogical knowledge; we are left with equivocal knowledge. This is precisely a pitfall Van Til wanted to avoid, and it is something analogical knowledge makes impossible. Of course, if such a definition does turn out univocal, then it is unclear why analogical knowledge is required in the first place, since at least one example of univocal knowledge has been provided; and that, precisely in the definition of analogical knowledge. If Van Til's theory of analogical knowledge is taken to its logical conclusion, therefore, Clark is correct that one can never really know anything.